Journal of Neurology

, Volume 253, Issue 9, pp 1197–1202

Dural sinus thrombosis in spontaneous intracranial hypotension

Hypotheses on possible mechanisms


  • Mario Savoiardo
    • Department of NeuroradiologyIstituto Nazionale Neurologico “C. Besta”
  • Silvia Armenise
    • Department of RadiologyCard. G. Panico Hospital
  • Pantaleo Spagnolo
    • Department of RadiologyF. Ferrari Hospital
  • Tiziana De Simone
    • Department of NeuroradiologyIstituto Nazionale Neurologico “C. Besta”
  • Maria Luisa Mandelli
    • Department of NeuroradiologyIstituto Nazionale Neurologico “C. Besta”
  • Alessandra Marcone
    • Department of NeurologySan Raffaele Hospital
  • Giancarlo Morciano
    • Department of RadiologyF. Ferrari Hospital
  • Cosma Andreula
    • Department of NeuroradiologyCittà di Lecce
  • Eliana Mea
    • Department of NeurologyIstituto Nazionale Neurologico “C. Besta”
  • Massimo Leone
    • Department of NeurologyIstituto Nazionale Neurologico “C. Besta”
    • Department of NeuroradiologyIstituto Nazionale Neurologico “C. Besta”

DOI: 10.1007/s00415-006-0194-z

Cite this article as:
Savoiardo, M., Armenise, S., Spagnolo, P. et al. J Neurol (2006) 253: 1197. doi:10.1007/s00415-006-0194-z


Dural sinus thrombosis (DST) is rarely associated with spontaneous intracranial hypotension (SIH). Engorgement of the venous system, caused by the CSF loss that occurs in SIH, is considered to favour the thrombosis, although signs of both SIH and DST are usually seen simultaneously at the first diagnostic MRI. We observed two patients with SIH and DST. Changes in pattern of headaches and MRI findings demonstrated that DST followed SIH. In SIH, the velocity of the blood flow in the dural sinuses may be reduced because of dilatation of the venous system which compensates the CSF loss. Other possible mechanisms seem unlikely on the grounds of both clinical presentation and MRI studies.


spontaneous intracranial hypotensionCSF lossdural sinus thrombosisvenous hypervolemiablood flow velocity


Dural sinus thrombosis (DST) has occasionally been observed in patients with spontaneous intracranial hypotension (SIH). The number of observations, although limited [4, 25], seems to indicate that their association cannot be simply interpreted as a coincidental finding.

SIH is considered to be always linked to CSF leakage, even if the point of CSF loss may remain cryptic to all neuroradiological or nuclear medicine investigations. In some cases, the pressure of CSF at lumbar puncture is within normal limits [19]; orthostatic headache, which is the commonest presentation of SIH, may be absent [18]. To encompass all atypical cases, therefore, Mokri [17] has emphasized that the most characteristic feature of the syndrome is not CSF hypotension but diminished CSF volume.

According to the Monro-Kellie doctrine [16], in a close compartment such as the intracranial and spinal dural space, any loss of one component must be compensated by an increase in another one. Therefore, to maintain the equilibrium of the compartment, the CSF volume that is lost must be replaced by an increase in the most easily expansible component, which is venous blood. In SIH, thickening of the pachymeninges, particularly demonstrated in post-contrast magnetic resonance imaging (MRI) by dural enhancement [9, 11, 20, 21], dilatation of the dural sinuses [2, 22, 23] and spinal epidural plexus [8], and enlargement of the pituitary gland [24] are the manifestations of the compensatory venous hypervolemia that are commonly observed on MRI studies.

In the reported cases of DST and SIH [4, 12, 25], it is not always clear which of the two happens first, DST or loss of CSF, because the features of both DST and SIH are usually found on the first MRI study and the sequence of events, particularly the possible changes in the characteristics of headache, is not always detailed [12].

We report two patients who presented with SIH and DST and discuss the possible mechanisms that may be involved in determining the association of these two manifestations.

Case reports

Patient 1

A 31-year-old man complained of sudden onset of cervical pain after exercising with a bar in a gym. A few days later, he complained of an increasingly severe fronto-occipital headache, followed by nausea and vomiting. The headache worsened in the upright position and improved when lying down. He presented at an emergency department where a brain CT scan was normal; the dural sinuses were large, with normal density. The following day, an MRI study was also normal.

The patient was given general analgesic, non-steroidal anti-inflammatory drugs and benzodiazepines with no benefit. Headache and vomiting persisted and blurred vision and hearing disturbances (tinnitus and hearing distortion) appeared. The headache worsened maintaining its positional characteristics, although less marked. The patient lost 10 kg of weight in 2 weeks and was admitted to hospital. Physical and neurological examinations were unremarkable. MRI showed thin subdural effusions, post-contrast enhancement of the thickened dura on both convexities and along the tentorium, and a marked sagging of the brainstem, consistent with SIH. MRI also showed signal abnormalities consistent with DST within the superior sagittal and left transverse and sigmoid sinuses (Fig. 1A). Cervical spinal MRI demonstrated enlargement of the epidural venous plexus. Bed rest, intravenous hydration, corticosteroids, and oral anticoagulant (warfarin) were started. Two weeks later, repeat brain MRI showed the evolution of the thrombus (Fig 1B). Spinal MRI, which included the upper thoracic segment, also showed an anterior epidural CSF collection from C5 to T2. The origin of the CSF leak was not demonstrated. In the following months, the headaches markedly improved and were no longer related to body position. Extensive investigation for thrombophilia was negative.
Fig. 1

Patient 1. Axial (A) T1-weighted post-contrast image shows thickened dura all over the convexity consistent with SIH. The intermediate signal within the superior sagittal sinus (arrowhead) indicates early non-enhancing thrombus. Midsagittal T1-weighted image without contrast (B) obtained 14 days later shows extensive hyperintensity of the thrombus due to methemoglobin formation. The straight sinus presents normal flow void. Sagging of the brainstem is also present

Follow-up brain MRI at 4 months no longer showed signs of SIH; signs of DST were improved with residual partial occlusion of the superior sagittal sinus and left transverse and sigmoid sinuses. Control MRI 18 months later showed nearly complete recanalization of the dural sinuses and no signs of SIH.

Patient 2

A 40-year-old woman presented with a 2-week history of paresthesias on the left side of her face and in both upper limbs, mild frontal headache not related to position and neck pain. Brain MRI showed bilateral subdural effusions or thickened pachymeninges over both convexities and along the tentorium, with high signal intensity in T2-weighted images, and diffuse dural enhancement after intravenous contrast medium administration. FLAIR and DW images demonstrated a small area of cortical hyperintensity in the upper part of the left frontoparietal convexity around the central sulcus consistent with a recent ischemic lesion. Phase contrast MR angiography showed normal veins and dural sinuses except for a questionably irregular short segment of the superior sagittal sinus at the vertex (Fig. 2A, B). Spinal MRI showed enlarged, enhancing epidural plexuses, particularly at the upper cervical level and at the thoraco-lumbar junction. Search for CSF leak, carried out with 3D “myelographic” sequences, was negative. Corticosteroid treatment was started and headache rapidly disappeared. However, the patient gradually became drowsy, with slowing down of speech and movement, and 10 days after MRI was admitted to the emergency department in a lethargic state. CT showed hyperdense dural sinuses without abnormal densities in the brain tissue. The following day, MRI confirmed the extensive thromboses of the superior sagittal sinus, the straight sinus and the initial segment of both transverse sinuses. The signal abnormalities previously seen on the left frontoparietal convexity had disappeared. However, signal hyperintensities, particularly evident in DWI, were present in both thalami, more marked on the left side, where they involved the posterior putamen, the posterior periventricular white matter and the parietal cortex, and extended caudally to the midbrain. Post-contrast enhancement of the thickened dura was reduced.
Fig. 2

Patient 2. First MRI: coronal FLAIR image (A) shows hyperintense effusions along the convexities and the tentorium (arrowheads) consistent with SIH. Focal cortical hyperintensity in the left upper rolandic area (arrow) is present. Venous MR angiography (B) is normal except for doubtful irregularities of the superior sagittal sinus at the vertex (arrowhead). Third MRI 18 days later shows near disappearance of signs of SIH, with hyperintense, thrombosed superior sagittal sinus (C, arrowhead). Venous MR angiography (D) shows extensive dural sinus thrombosis (arrowheads). The posterior part of the great vein of Galen and initial straight sinus are thrombosed (asterisk)

The patient was treated by intravenous heparin infusion with rapid improvement. Only a mild, prevalently orthostatic, occipital headache persisted. Extensive investigation for thrombophilia was negative. One week later, control MRI showed persistent thrombosis of the dural sinuses, but regression of the signal abnormalities in the right thalamus and left parietal cortex. The other areas involved showed decreased signal abnormalities (Fig. 2C, D). The thin pachymeningeal enhancement was unchanged. After two weeks, the patient became asymptomatic. The last MRI follow-up, 6 months after onset, shows extensive but incomplete recanalization of the superior sagittal sinus and complete recanalization of the other dural sinuses; no signs of SIH are present. The patient remains asymptomatic, still on oral anticoagulants.


In both our patients, the sequence of events at the onset of the disease seems to favor the development first of SIH, soon followed by DST. In the first patient, this impression is mainly supported by the changing pattern of his headache. In the second one, by the signs suggesting SIH on the first MRI study, in which, however, questionable abnormalities of the superior sagittal sinus were also seen. The focal cortical abnormalities in the left frontoparietal cortex might indicate ischemic edema from stagnation due to thrombosis or slow flow in a parasagittal cortical vein. Extensive DST occurred later and was demonstrated only by the second MRI. In both patients, spinal MRI showed collapse of the dural sac with dilatation of the epidural venous plexuses, consistent with SIH. To our knowledge, epidural spinal plexus engorgement has never been reported in intracranial DST, and, in SIH, is considered a compensatory sign for diminished volume of CSF [14]. Although CSF pressure measurements were not obtained, in our cases, presence of epidural plexus engorgement (and epidural CSF collection in patient 1) further supports the diagnosis of SIH as the presenting condition. The current hypothesis for explaining DST in SIH is that the engorgement of the venous system due to the compensatory venous hypervolemia may favor the thrombosis within the dural sinuses [4, 25]. This hypothesis is very reasonable and certainly the cerebral veins and dural sinuses often appear markedly enlarged in SIH. Assessment of the enlargement, however, is usually difficult because of the large variation in the size of the dural sinuses in normal conditions [7]. Recognition of the enlargement of the dural sinuses is obvious when the MR images obtained during SIH are compared with those obtained after disappearance of SIH, either spontaneous or following epidural blood patch or other therapeutic interventions.

Insights into the pathogenetic mechanism leading a patient with decreased intracranial pressure to DST are offered by the observation of cases with DST following lumbar puncture. Cases of SIH and DST are rare [4, 12, 25], whereas patients with DST following lumbar puncture have been more frequently reported [1, 3, 5, 6, 10, 12, 13, 26]. In these patients, there is no question of which condition happens first; CSF hypotension caused by the lumbar puncture obviously precedes the venous thrombosis. The pathogenetic mechanisms leading to DST should not be different between the patients with SIH and those submitted to lumbar puncture or other epi-intradural injections.

How relevant the slowing of blood flow associated with the dilatation of the dural sinuses in favoring DST might be, is unknown. In steady flow, the fluid velocity through a section is inversely proportional to the size of the section. Therefore, dilatation of the sinuses or of any vessel correlates with a proportional decrease in the blood flow velocity, if all the other conditions are unchanged. However, it is not easy to determine the decrease in blood flow velocity that occurs in the sinuses of patients with SIH. Canhão et al. [7], considering the patients who had been reported to develop DST after lumbar puncture, investigated the possible mechanisms involved by measuring the blood flow velocity in the straight sinus of patients before and after lumbar puncture using transcranial Doppler ultrasound. They found a 47% decrease in mean blood flow velocity that persisted for a few hours. In SIH patients, in whom the loss of CSF volume is probably greater than that of the patients who had been submitted to a lumbar puncture, it is reasonable to suppose that the slowing of the blood flow is even more marked.

It is worth noting that, in spite of the name “dura” which means “hard”, the dural walls of the sinuses are not rigid. As we mentioned above, observations of cases of SIH have demonstrated that the dural sinuses do vary in size in response to variations in CSF volume and pressure.

To quantify the modifications of the dural sinuses in the two different conditions, i.e., presence of SIH (SIH+), and clearing of SIH, with the patient in normal clinical condition and with normal MRI (SIH-), we reviewed three patients whose SIH+ and SIH- MRI slices matched perfectly and were, therefore, comparable (Fig. 3). The measurements performed on 5 sections demonstrated that the areas of the dural sinus sections in SIH+ were about 70% larger than the corresponding sections in SIH- condition, i.e., the normal condition (range 39%–98%, average: 69.8%). If all other conditions were unchanged or not influencing the blood velocity, this change would indicate that the slowing of the blood flow is remarkable: in a few cases, the velocity might almost be half of the blood flow velocity in normal conditions. If some risk factors for thrombosis were present, as reported in a few cases of the literature [4, 13, 25] such a slowing could very well be a concurrent cause for DST. In our patients, no blood abnormalities favoring thrombosis were found. The reason why some people tolerate a slow flow without developing thrombosis is unknown.
Fig. 3

Sagittal proton density midline sections in a patient in SIH+ (A) and SIH- (B) conditions. In SIH+ there is marked dilatation of the straight sinus (arrow); the internal cerebral vein and the inferior vermian vein are also larger than in normal conditions (SIH-), ( arrowheads). Axial T2-weighted sections in a different patient in SIH+ (C) and SIH- condition (D) at the same level show marked enlargement of the superior sagittal sinus (arrowheads) and bilateral subdural effusions in SIH+. Areas of the sections of the sinuses were calculated after manual segmentation: ratio of C/D is 198/100

To our knowledge, the alternative sequence of events, i.e., DST followed by or causing SIH, has never been demonstrated and is more difficult to explain. One might suppose that the increased intracranial pressure caused by DST would break the dura at a congenitally or acquired weak point (for instance, a Tarlov’s cyst), thus causing CSF leakage and a secondary intracranial hypotension. Although such a mechanism might not be excluded, we could not find any evidence from the cases reported in the literature that the changes in the pattern of headache or MRI studies would favor this sequence. The sequence, almost constantly reported both in SIH patients and in patients with post-lumbar puncture CSF hypotension developing DST, is that of a changing pattern of headache: the orthostatic characteristics rapidly diminish, giving way to a more severe headache which is no longer relieved by lying down or may even be aggravated by recumbency. This change in the pattern of headache should alert the physician about the possible thrombotic complication [1].

A third hypothesis could be considered: the thickening of the pachymeninges observed in our cases could be the expression of dural veins engorgement due to DST. If this were true, the caudal displacement of the brain should not be interpreted as sinking of the brain caused by CSF loss, but would be the result of a downward pressure or mass effect caused by the diffusely thickened dura. This interpretation seems very unlikely for at least three reasons: 1) such a phenomenon has never been observed in isolated, even extensive thrombosis of the dural sinuses; 2) spinal signs (dilated epidural plexus, collapse of the dural sac, and CSF epidural collections) militate against this hypothesis; 3) orthostatic headache relieved by recumbent position (as in our patient 1 and in Berroir’s and Sopelana’s cases) [4, 25] is well explained by intracranial hypotension and not by hypertension associated with DST.

In conclusion, in SIH, venous hypervolemia is present to compensate for the CSF diminished volume. The velocity of blood in the dural sinuses is diminished and such slowing may favor thrombosis. Therefore, in SIH patients it is confusing to use the term “CSF hypovolemia” that has been introduced to encompass the cases in which CSF pressure is found normal [14, 15]. This term is incorrect because the suffix –emia indicates blood, not CSF. Simple, descriptive expressions, such as “decreased volume of CSF”, should be preferred.


We thank Dr. Nicoletta Anzalone, San Raffaele Hospital, for the MR images of case 1.

Copyright information

© Steinkopff Verlag Darmstadt 2006